Temperature-dependent thermal properties for cross-laminated timber exposed to standard fire

Abstract

In the draft of the Eurocode 5‐1‐2 revision (2020), the charred depth of CLT is modelled using a simplified model, which takes into account two different scenarios: linear charring when the bond line integrity is maintained and stepwise bilinear charring when the bond line integrity is not maintained. That means that fall‐off of charred CLT layers is taken into account when during fire exposure the bond line integrity is not maintained. Numerical simulations based on FE heat transfer models are usually conducted based on the temperature‐dependent thermal properties for wood and the char layer according to the current Eurocode 5‐1‐2 (2004). König and Walleij (1999) proposed these thermal properties for initially unprotected and protected timber members. Thus, fall‐off has been considered as single incident in the case of fall‐off of the fire protection system. However, it is questionable if the thermal properties according to the current Eurocode 5‐1‐2 (2004) can be used for CLT plates showing multiple fall‐off of charred layers. In the frame of an ongoing research project the fire behaviour and fire resistance of cross‐laminated timber rib panels (CLT rib panels) is studied with experimental and numerical analysis. The floor system consists of cross‐laminated timber plates rigidly bonded to glued‐laminated timber ribs by means of screw‐press gluing. Two different types of cross‐section are studied. The experimental results of a series of full‐scale fire tests recently concluded showed more than 90 and 120 min of fire resistance and confirmed the assumption of fall‐off of single or multiple charred CLT layers. The proposed paper focuses on the numerical simulation of the charring behaviour of the CLT plates exposed to ISO‐standard fire on one side, taking into account multiple fall‐off of CLT layers. The proposed paper provides the results of numerical simulations, which were conducted for the analysis of the development and distribution of the temperature within the four tested cross‐sections, in comparison with the experimental results and the estimated results according to the simplified design method of Eurocode 5‐1‐2 revision (2020). For this objective, FE 2D heat transfer models were generated modelling the cross‐section of a T‐section or box‐section. Fall‐off of charred CLT layers were defined as time steps of the thermal simulation when the average temperature between CLT layers exceeded 300°C. In a further step, the FE model was re‐created omitting the fallen‐off layer and the calculation was continued until the next layer has fallen off. Using the temperature‐dependent thermal properties according to Eurocode 5‐1‐2 (2004), the thermal simulations predicted a significantly earlier fall‐off of the charred CLT layers in comparison with the experimental results. The proposed paper presents a new set of thermal properties calibrated to the experimental results of the tested cross‐sections of the CLT plates. The validation of the calibrated data is made by comparing additional thermal simulations with further experimental results (mostly single CLT plates). This leads to the proposal to use the new set of temperature‐dependent thermal properties for cross‐laminated timber to consider the effect of multiple fall‐off of charred layers.